Stereoscopic player observation
 

I have noticed that the quality of the 3D output when playing raw left and right video files is noticeably higher then the same video after rendering to frame compatible side by side format even if it's rendered to an uncompressed AVI.
This is observed when playing both with Stereoscopic player outputting in checkerboard format to a Samsung UN40C7000 3DTV.
Shouldn't the quality be the same?
After all they are both half res.
The video was acquired with dual Canon HF21s using MXP mode at 60i.

you said yourself, after rendering.
it doesn't mean because you save uncompressed video that the process has not damaged the data in between. Jpeg, mpeg2 and mpeg4 are very sensitive to resizing and rotating picture.
And possibly you used an application/profile that works "lossy" using to few bits or the wrong format (4:2:0 instead 4:2:2)

I guess I should have been more specific.
I used Sony Vegas 10ver-c and without modifying in any way,
rendered it in half res side by side format using the Microsoft Codec uncompressed.
The reduction in image quality is very noticeable which surprised me.

No. The checkerboard format combines the pixels from the two images like this:

L1, R2, L3, R4, ...
R1, L2, R3, L4, ...
etc.

The side-by-side half-resolution combines them differently:

L1 + L2 / 2, L3 + L4 / 2, ... | R1 + R2 / 2, R3 + R4 / 2, ...

So the result in quality can be quite different.

I am displaying them both using checkerboard output.

Yes, but in one case your original is not compressed and in the other it is.

I see what is happening now.
If I play the left and right files using Stereoscopic Player in checkerboard mode I get a very good image.
If I play the unaltered rendered sbs version in Stereoscopic Player in checkerboard mode, I get noticeably lower quality.
If I play the rendered sbs version with another media player and force the TV into side by side 3D mode the image quality is good again.
Apparently playing half res sbs using checkerboard cuts the resolution once again.
Interesting.